Abstract: Objective To find biomarkers and establish a serum protein fingerprint model for early diagnosis of ESCC through matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS). Methods Serum samples were collected from 75 patients with ESCC and 44 healthy individuals.Proteomic spectra of mass to charge ratio (m/z) was generated by weak cationic-exchanger magnetic beads (WCX-MB).Data were processed by protein microarray data analysis system.A diagnosis model to identify normal from early esophageal cancer or advanced esophageal cancer was established using support vector machines combined with genetic algorithms.79 specimens (50 cases of esophageal cancer and 29 cases of healthy controls) were randomly selected for training and cross-validation.At last，the new (30 patients with esophageal cancer and 23 healthy controls) serum specimens were detected as a testing cohort. Results Two different patterns were established by MALDI-TOF MS.Pattern 1，using 11 protein peaks，were able to separate ESCC patients from the healthy individuals with a sensitivity of 92.4% and a specificity of 87.4%.Pattern 2，using 8 protein peaks，were able to separate ESCC of stage I and stage II from stage III and stage IV with a sensitivity of 97.5% and a specificity of 85.7%. Conclusion These results suggested that MALDI-TOF MS combined with MB separation could generate a serum protein fingerprint model with high sensitivity and specificity for ESCC diagnosis.